MUTLA: A Large-Scale Dataset for Multimodal Teaching and Learning Analytics
Project Overview
The document discusses the utilization of generative AI in education, emphasizing the MUTLA dataset, which comprises a comprehensive collection of multimodal data, including learning logs, videos, and EEG brainwave information from students engaged with the Squirrel AI Learning System. This dataset serves as a pivotal resource for researchers aiming to improve teaching and learning analytics by analyzing authentic educational interactions and predicting student engagement levels. Such insights are crucial for refining adaptive learning systems, ultimately leading to enhanced educational outcomes. The findings highlight the potential of generative AI to personalize learning experiences, adapt instructional strategies based on individual student needs, and foster greater engagement, thus transforming the educational landscape. By leveraging advanced AI capabilities, educators can better understand student behaviors and learning patterns, making informed decisions that support effective teaching and learning processes. Overall, the integration of generative AI and the insights derived from the MUTLA dataset signify a promising advancement in the quest for more effective and responsive educational environments.
Key Applications
MUTLA dataset for multimodal teaching and learning analytics
Context: After-school learning centers in China, targeting primary and middle school students
Implementation: Data collected from students using Squirrel AI Learning products, including learning logs, EEG data, and video recordings during tutoring sessions.
Outcomes: Improved understanding of student engagement through analysis of multimodal data, which can enhance adaptive learning systems and educational outcomes.
Challenges: Limited accessibility to high-quality multimodal data for teaching and learning analytics.
Implementation Barriers
Data Quality
Teaching and learning analytics is limited by the quality and quantity of multimodal data that is publicly accessible.
Proposed Solutions: Providing a publicly accessible large-scale dataset like MUTLA to bridge the gap between researchers and data.
Project Team
Fangli Xu
Researcher
Lingfei Wu
Researcher
KP Thai
Researcher
Carol Hsu
Researcher
Wei Wang
Researcher
Richard Tong
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Fangli Xu, Lingfei Wu, KP Thai, Carol Hsu, Wei Wang, Richard Tong
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai